Optical networks using ROADMs for routing optical wavelengths are becoming increasingly prevalent resulting from the need for lower costs, greater flexibility between data formats, and efficient capacity upgrades. In optical networks, the achievable spectral efficiency (SE) and overall fiber capacity are limited by noise from optical amplifiers and fiber nonlinearity. The SE and overall fiber capacity are also limited by channel bandwidth narrowing caused by the usage of wavelength selective devices or wavelength selective switches (WSS), which are components of ROADMs.
For example, in 50 GHz-spaced 40-Gb/s and 100-GB/s wavelength-division-multiplexed (WDM) systems, channel bandwidth narrowing effects are managed by using digital coherent detection and bandwidth improvement of wavelength selective devices. However, even with combined use of digital coherent detection and improved flat-top wavelength selective devices, the usable bandwidth for a typical long-haul optical network is still limited to 30-40 Ghz for a 50 Ghz-spaced WDM system.
Usable bandwidth reduces as the number of cascaded ROADMs increases. As an example, if there are 20 cascaded ROADMs, the usable bandwidth is typically limited to about 30 GHz. Although this channel bandwidth supports 100 Gb/s systems over 50 GHz WDM grid by using polarization-multiplexed (PM) quadrature phase shift keying (QPSK) modulation, it is not wide enough to support future 400 Gb/s and beyond systems operating at higher spectral efficiency by using high-order quadrature amplitude modulation (QAM) based modulation formats.
Improvements to spectral utilization in WDM systems have been proposed. A first method is based on the concept of using a super channel where the channel grid is increased in order to reduce the portion of channel guard bands to improve spectral utilization. However, usage of the super channel reduces channel granularity, and thus reduces network efficiency and flexibility. A second method is based on using transmitter side pre-equalization, where the spectral shape of a transmitted signal is pre-emphasized in order to combat cascaded optical filtering along optical links. However, because the narrow optical filtering effects occur in a distributed way along each optical link, the pre-equalized signal will have a higher launch power and as a result experience more nonlinear impairments. A third method utilized advanced post equalization techniques such as maximum likelihood sequence equalization (MLSE) to perform post-transmission equalization of filtering effects. However, this method does not perform well when the signal-to-noise ratio is low, which is typical for high-speed systems using advanced feed-forward error correction coding.